Merck Life Science: Ensuring scientists get the products they need with powerful ML models
About Merck Life Science
Life Science business of Merck KGaA Darmstadt, Germany provides scientists with the state-of-the-art tools, services, and expertise they need to perform experiments and engineer new products. Offering one of the broadest portfolios in the industry, along with best-in-class products for pharmaceutical development and manufacturing, Merck Life Science accelerates access to health for people everywhere.
Tell us your challenge. We're here to help.
Contact usLife Science business of Merck KGaA Darmstadt, Germany moved its data stack to Google Cloud, giving it powerful insights to reduce inefficiencies in its supply chains and help customers find the right products at the right time.
Google Cloud results
- Increases traffic to product-comparison page by 400% with similar-product grid built in Vertex AI
- Improves data literacy across the organization, driving 650% more traffic to Looker ecommerce dashboards in just two years
- Builds product recommendation models and removes inefficiencies in supply chain with BigQuery and Vertex AI
Looker ecommerce dashboard traffic up 650% in two years
For the life science industry, coordinating product supply and demand can be a challenge. Factors such as the post-pandemic mRNA boom or the fast ramp up and down of major lab projects create volatile demand, which can be difficult for businesses to meet smoothly. This is compounded by the vulnerability of globalized, complex supply chains, which make it harder to maintain a constant supply of products.
Selling a wide spectrum of products to buyers across the life science industry, from pharmaceutical companies to food and drinks manufacturers, Life Science business of Merck KGaA Darmstadt, Germany is particularly exposed to these challenges. To make the right products available at the right time to the customers that need them, Merck Life Science needs to ensure its supply lines are efficient, as well as making it easy for buyers to find what they are looking for. This requires departments across the business being able to use data to gain insights into supply chains and demand fluctuations and respond quickly to market needs. But until recently, Merck Life Science was limited by the capabilities of its third-party data solution, which could only offer limited insights, such as the number of page views or product sales.
"Descriptive stats are fine, but we needed to go much deeper, to be able to build algorithms and models to drive predictions, forecasting, anomaly analysis, and recommendations in real time," explains Gail Buffington, Head of Data Science and Analytics, Merck Life Science. "It was at that point that we realized that we needed to make a change."
Finding the simple solution to complex data sets
In 2021, Merck Life Science was growing rapidly, and Buffington was building her team to keep pace. Keen to enable the team to make the most of its business data, Buffington migrated the company's data stack to Google Cloud. Not only did this allow for a straightforward integration with Google Analytics, which Merck Life Science was already using, Buffington also appreciated the ease of use of Google Cloud.
"We wanted something accessible and approachable," says Buffington. "Google Cloud is probably the best-documented system out there, it's simple to use, and I can train a team member to use it very quickly. Our data is already complex. I didn't want to add complexity with our data solution."
"We wanted something accessible and approachable. Google Cloud is probably the best-documented system out there, it's simple to use, and I can train a team member to use it very quickly. Our data is already complex. I didn't want to add complexity with our data solution."
—Gail Buffington, Head of Data Science and Analytics, Merck Life ScienceHelping customers find what they need with relevant product recommendations
With a range of hundreds of thousands of highly diverse products on its site, Merck Life Science knew that it could be difficult for customers to find what they needed. It wanted to reduce the amount of friction customers experienced, ensuring they could find and purchase products easily. With all its Google Analytics data now combined with the data from its enterprise resource planning (ERP) platform in BigQuery, Merck Life Science is able to use Vertex AI to build buy-it-again product recommendation models. It also builds personalized product recommendation models for individual users, which take into account the fact that some users (say in food manufacturing) use the same products in very different ways to others (such as in pharmaceuticals).
"These product recommendations are a huge revenue driver for us, making a very complex website significantly more customer friendly," says Buffington. "We could have done it using the previous setup, but it would have been extremely hard, and every time we wanted to make a change, it would have been complex. Whereas now we release some sort of enhancement, a new placement, or new recommendations every other week. It's a constant stream of releases."
Merck Life Science has also helped customers to distinguish quickly between similar products by using algorithms to highlight key differences in a comparison grid. This helps the customer to identify at a glance the product they need, rather than having to jump from product page to product page, as they were doing previously. Those customers who need more information are then driven to the product comparison pages, with traffic to these pages now five times higher than it was before the similar-product grids were introduced.
Building efficient supply chains with machine-learning models
Selling such a vast range of products, each with their own supply chain, creates significant complexity in trying to ensure products are in the right place at the right time for customers. Coordinating global warehouse stocks with regional demand, ensuring manufacturing centers contribute their parts promptly, making sure that goods are delivered to meet customers' delivery expectations: coordinating these many moving parts is a significant challenge. With the previous system, this challenge was compounded by the fact that transforming and moving data from the ERP to the third-party data solution took around two days to complete, meaning that its data was already two days behind events.
Now, Merck Life Science merges all its complex data sets in BigQuery to gain insights in near real time, before building machine-learning models in Vertex AI to enable its supply lines to meet the fluctuating demands of its global customer base.
"Having excess inventory, or not enough to meet demand, affects our bottom line, as can operating inefficient shipping routes. Building models to meet these challenges involves vast reams of data," Buffington explains. "Google Cloud allows us to process, clean up, and document data much more quickly, enabling us to meet customer delivery expectations and reduce inefficiencies throughout our supply chain."
"Google Cloud allows us to process, clean up, and document data much more quickly, enabling us to meet customer delivery expectations and reduce inefficiencies throughout our supply chain."
—Gail Buffington, Head of Data Science and Analytics, Merck Life SciencePutting data in the hands of the people who need it
Buffington's team has helped to foster a data culture throughout Merck Life Science, using Looker dashboards to democratize access to business data across departments, including customer experience, digital marketing, procurement, and operations. Ecommerce dashboards in Looker receive 650% more traffic than they did just two years ago, indicating a significant increase in data literacy and transparency across the company. As Buffington explains, "Google Cloud has enabled us to gain much faster insights, and become a much more data-driven organization, making departments comfortable asking my team to help them reach increasingly complex data insights."
For example, by combining a number of site-page engagement metrics such as time on page, scroll depth, bounce rate, and exit rate, Buffington's team has built a Looker dashboard that is able to score each page on how engaging it is for customers who are researching products. These scores are then used to optimize the pages for greater engagement, helping more customers find the products they want more quickly.
Always innovating to meet customer expectations
Merck Life Science is currently looking at how it can further develop its Looker dashboards using generative AI to offer users more insights into product demand. One project in development is to build dashboards that use generative AI to explain why a certain product has seen a sudden surge in purchases. Users would be able to go into the dashboard, spot a spike in demand and query it, with generative AI providing context such as explaining that the buyer recently received a large funding award, or began a particular project. In this way, users will be able to understand the needs of customers more clearly, helping to ensure a smooth supply of products.
Merck Life Science is also using generative AI to gain a deeper understanding of customer feedback. Users can ask natural-language questions about customers' issues with particular products, and the generative AI tool searches through all customer reviews to deliver clear summaries of how customers feel about products, saving the business time having to trawl through data to understand customers' concerns.
As Merck Life Science grows, Buffington continues to monitor the Google Cloud release calendar to ensure it is able to continue getting the right products to its customers at the right time. "Our roadmap is very driven by what Google Cloud is releasing," she says. "We're constantly excited by solutions being rolled out, and we devote a lot of time to exploring new features from Google Cloud to help us meet customers' expectations long into the future."
"Our roadmap is very driven by what Google Cloud is releasing. We're constantly excited by solutions being rolled out and we devote a lot of time to exploring new features from Google Cloud to help us meet customers' expectations long into the future."
—Gail Buffington, Head of Data Science and Analytics, Merck Life ScienceTell us your challenge. We're here to help.
Contact us
About Merck Life Science
Life Science business of Merck KGaA Darmstadt, Germany provides scientists with the state-of-the-art tools, services, and expertise they need to perform experiments and engineer new products. Offering one of the broadest portfolios in the industry, along with best-in-class products for pharmaceutical development and manufacturing, Merck Life Science accelerates access to health for people everywhere.